Abstract:

Utilizing daily meteorological data from 17 basic meteorological observation stations, along with potato yield and growth period data in Hebei province from 1982 to 2022, we compared potato yield of separation calculation using methods of three-year moving average, five-year moving average, five-point quadratic smoothing, Hodrick-Prescott filtering, and quadratic exponential smoothing, and applied the key meteorological factor method and the climate suitability method to simulate and validate potato yield, analyzing the correlation coefficients, root mean square errors, and forecast accuracy between the fitted and actual yields. The results indicated that all methods can fit the yield trend well. Meteorological yield exhibits significant inter-annual variability, with considerable differences among the yield separated by the different methods. For yield forecasting, the quadratic exponential smoothing method performs the best, followed by the Hodrick-Prescott filtering method, while the 3-year moving average method is the least effective.

Key words: Yield separation, Trend yield, Meteorological yield, Yield forecasting